MathDB2 problems
- 0 votes0 replies0 views
Polynomial-bounded convergence error for continuous policy optimization
Polynomial-bounded convergence conjecture. The convergence rate is likely to be polynomial-bounded under mild assumptions, extending beyond the condition required by the cited anal…
- 0 votes0 replies0 views
Improved natural-gradient complexity for LQR
Natural-gradient complexity conjecture. This bound should be improvable at least to